# Poisoning Defense in U-shape Split Learning on ImageNette

Source: [Poisoning Defense in U-shape Split Learning on ImageNette](<https://www.sota2.com/research/sota/poisoning-defense-in-u-shape-split-learning-on-imagenette>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:31:17.222Z

Evaluation of defense mechanisms against poisoning attacks specifically within a U-shape Split Learning framework applied to the ImageNette dataset, reporting accuracy (ACC) and attack success rate (ASR).

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Poisoning Defense in U-shape Split Learning |
| Dataset (including any stated split) | ImageNette |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Accuracy](<https://www.sota2.com/research/sota/poisoning-defense-in-u-shape-split-learning-on-imagenette.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Attack Success Rate (ASR)](<https://www.sota2.com/research/sota/poisoning-defense-in-u-shape-split-learning-on-imagenette.md?metric=Attack%20Success%20Rate%20(ASR)>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| SecureSplit | 64 | Not recorded | 69705dcce25ad2fa9d1ed601 |
| No defense | 62 | Not recorded | 69705dcce25ad2fa9d1ed601 |
| ANP | 61 | Not recorded | 69705dcce25ad2fa9d1ed601 |
| DP | 60 | Not recorded | 69705dcce25ad2fa9d1ed601 |
| SafeSplit | 60 | Not recorded | 69705dcce25ad2fa9d1ed601 |
| MP | 59 | Not recorded | 69705dcce25ad2fa9d1ed601 |
| TrMean | 57 | Not recorded | 69705dcce25ad2fa9d1ed601 |
| HDBSCAN | 56 | Not recorded | 69705dcce25ad2fa9d1ed601 |
| VFLIP | 53 | Not recorded | 69705dcce25ad2fa9d1ed601 |
| Multi-Krum | 52 | Not recorded | 69705dcce25ad2fa9d1ed601 |

### SecureSplit

Accuracy: 64. Result ID: 6996e9e45698665ec4365c40.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

Recorded evaluation configuration: {"Defense":"SecureSplit"}. This record does not assert independent reproduction.

### No defense

Accuracy: 62. Result ID: 6996e9e45698665ec4365c37.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

Recorded evaluation configuration: {"Defense":"None"}. This record does not assert independent reproduction.

### ANP

Accuracy: 61. Result ID: 6996e9e45698665ec4365c3d.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b181e209131cea2c1a4.md>)

[Implementation](<https://github.com/csdongxian/ANP_backdoor>)

Recorded evaluation configuration: {"Defense":"ANP"}. This record does not assert independent reproduction.

### DP

Accuracy: 60. Result ID: 6996e9e45698665ec4365c3b.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65301e209131cea08bf5.md>)

Recorded evaluation configuration: {"Defense":"DP"}. This record does not assert independent reproduction.

### SafeSplit

Accuracy: 60. Result ID: 6996e9e45698665ec4365c3f.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

Recorded evaluation configuration: {"Defense":"SafeSplit"}. This record does not assert independent reproduction.

### MP

Accuracy: 59. Result ID: 6996e9e45698665ec4365c3c.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa646b1e209131cea04983.md>)

Recorded evaluation configuration: {"Defense":"MP"}. This record does not assert independent reproduction.

### TrMean

Accuracy: 57. Result ID: 6996e9e45698665ec4365c38.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69531e209131cea218fb.md>)

Recorded evaluation configuration: {"Defense":"TrMean"}. This record does not assert independent reproduction.

### HDBSCAN

Accuracy: 56. Result ID: 6996e9e45698665ec4365c3a.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69141e209131cea2013b.md>)

Recorded evaluation configuration: {"Defense":"HDBSCAN"}. This record does not assert independent reproduction.

### VFLIP

Accuracy: 53. Result ID: 6996e9e45698665ec4365c3e.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa749f1e209131cea62ad5.md>)

Recorded evaluation configuration: {"Defense":"VFLIP"}. This record does not assert independent reproduction.

### Multi-Krum

Accuracy: 52. Result ID: 6996e9e45698665ec4365c39.

[Reporting paper](<https://www.sota2.com/research/paper/69705dcce25ad2fa9d1ed601.md>)

Recorded evaluation configuration: {"Defense":"Multi-Krum"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.